The EHR IM specification is not sufficiently clear on the meaning of the times in the
OBSERVATION.data History class. The key issue is that the time points/periods for when
observations are recorded are understood as the times at which the data were true in
the subject of the observation (the patient). To be more precise, we differentiate
between 'time of sampling' and 'time of observing the sampled data'. These are almost
always the same (e.g. when taking a pulse the time the nurse or instrument generates
"90bpm" is within milliseconds of the time when the relevant finger or transducer did
However, in a few cases, like microbiology, a tissue sample may be taken from the body
and then the observation (the microscopy) done miinutes or hours later. The sampling time
is therefore different from the observation time.
The problem is to decide what the meaning of Observation.data.origin etc is. The important
point is that the observed values (no matter when they are observed) correspond to a state
of the patient's body at the time of sampling. There is a tension here: in normal science
if a certain set of data are seen, it is imperative to record exactly when those
values were seen.
WHat is needed here is a) a standardised meaning for Observation.data.origin in all cases,
and b)an ability to record a second time or offset if needed for cases when the sampling
and observation times are different.